arxiv: 2605.13120 · v1 · submitted 2026-05-13 · 📡 eess.SY · cs.SY· math.OC

Recognition: 2 theorem links

· Lean Theorem

D-Optimized Sampling Design for System Identification

Enrico Dozzi, Rodrigo A. Gonz\'alez, Tom Oomen

Authors on Pith no claims yet

Pith reviewed 2026-05-14 18:57 UTC · model grok-4.3

classification 📡 eess.SY cs.SYmath.OC

keywords system identificationnonparametric estimationfrequency response functionnonuniform samplingD-optimalitymultisine excitationspectral leakage

0 comments

The pith

Irregular sampling times optimized for D-optimality improve the accuracy of nonparametric frequency response estimates under nonperiodic multisine excitation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a nonparametric frequency response estimator that works with nonperiodic multisine inputs and nonuniform sampling times. It then designs the sampling instants to maximize the determinant of the information matrix, which directly reduces the variance of the resulting estimates. This approach targets situations where periodic excitation and uniform sampling cannot be used. A sympathetic reader would care because it extends reliable frequency-domain identification to more practical, irregular measurement conditions without requiring signal periodicity.

Core claim

By selecting nonuniform sampling times that maximize the determinant of the Fisher information matrix associated with the nonparametric frequency response estimator, the variance of the estimates is reduced and spectral leakage is mitigated for nonperiodic multisine inputs.

What carries the argument

D-optimal design of nonuniform sampling instants that maximizes the determinant of the information matrix for the leakage-aware nonparametric FRF estimator.

If this is right

System identification becomes feasible in applications where only irregular sampling is available.
The estimator covariance can be shaped directly through choice of sampling instants rather than input redesign.
Leakage effects in the frequency domain are controlled by sampling design even without periodic signals.
The method yields explicit expressions for the asymptotic variance that can be minimized offline.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same sampling design principle could be applied to other nonparametric estimators such as impulse response or transfer function models.
In closed-loop experiments with timing jitter, the D-optimized points might be adapted online to maintain information content.
Extension to multi-input multi-output systems would require a matrix-valued version of the D-criterion on the sampling schedule.

Load-bearing premise

The nonparametric frequency response estimator stays unbiased and leakage-free when the input is nonperiodic and the sampling times are nonuniform.

What would settle it

If the empirical covariance of the frequency response estimates obtained with the proposed sampling times is not smaller than that obtained with uniform sampling under identical nonperiodic multisine excitation, the claimed improvement does not hold.

Figures

Figures reproduced from arXiv: 2605.13120 by Enrico Dozzi, Rodrigo A. Gonz\'alez, Tom Oomen.

**Figure 2.** Figure 2: Information density of the least-squares FRF es [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

read the original abstract

Traditional system identification with multisine inputs relies on uniform sampling and periodic excitation to preserve Fourier orthogonality and avoid spectral leakage, limiting its use in scenarios with irregular sampling or nonperiodic inputs. This work investigates continuous-time system identification under nonperiodic multisine excitation and nonuniform sampling. We develop a nonparametric frequency response function estimator suited to such conditions and design irregular sampling schemes that enhance the informativeness of measurements and reduce spectral leakage. The proposed sampling scheme improve the statistical accuracy of system identification in settings where periodic excitation is impractical.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper introduces a nonparametric FRF estimator and D-optimal irregular sampling for nonperiodic multisine system ID, but the unbiasedness under nonuniform sampling rests on an unshown derivation.

read the letter

The main takeaway is that this work targets a practical gap: extending multisine system identification beyond the usual periodic uniform sampling by building a nonparametric frequency response estimator for nonperiodic inputs and nonuniform grids, then using D-optimality to pick sampling locations that improve measurement quality and cut leakage. That combination is the actual new piece relative to the standard literature on periodic excitations. They frame the limitation clearly and the direction makes sense for applications where periodic forcing is hard to arrange. The estimator and design are presented as a joint solution, which is a reasonable way to approach it. The soft spot is exactly the one the stress-test flags. The abstract claims the estimator stays unbiased and leakage-free, but gives no derivation, bias bound, or orthogonality argument that survives arbitrary sampling irregularity. Standard leakage cancellation leans on periodicity and uniform spacing; if those are dropped without a replacement proof that bias terms do not grow with irregularity, then optimizing the points cannot reliably deliver better accuracy. The full paper would need to show that math explicitly, plus some validation data, before the central claim holds up. This is for control engineers and system-ID people who routinely deal with irregular sampling or nonperiodic test signals. A reader already working on nonparametric FRF methods or experimental design might pick up the estimator idea or the optimization framing, but only after the derivations are checked. It deserves peer review so the bias analysis can be examined properly; the idea is grounded enough to warrant referee time even if revisions on the proofs are needed.

Referee Report

2 major / 1 minor

Summary. The paper develops a nonparametric frequency response function (FRF) estimator for continuous-time system identification under nonperiodic multisine excitation and nonuniform sampling. It further proposes D-optimal irregular sampling schemes intended to increase measurement informativeness and mitigate spectral leakage, with the claim that these designs improve statistical accuracy of system identification in applications where periodic excitation is impractical.

Significance. If the estimator is shown to be unbiased and leakage-free under the stated conditions, and if the D-optimized sampling demonstrably reduces variance without introducing bias, the work would extend classical frequency-domain identification to a broader class of practical measurement settings (embedded systems, event-triggered sampling, etc.). The explicit use of D-optimality on sampling locations is a potentially useful contribution provided the underlying estimator properties are rigorously established.

major comments (2)

[Abstract] Abstract: the nonparametric FRF estimator is asserted to remain unbiased and leakage-free for nonperiodic inputs and nonuniform sampling, yet no derivation, bias expression, or leakage bound is supplied. Standard leakage cancellation relies on periodicity and uniform grids; without an explicit proof that residual bias terms remain independent of sampling irregularity, the subsequent D-optimization of sampling points cannot be guaranteed to improve statistical accuracy.
[Abstract] Abstract: the central claim that the proposed sampling scheme improves statistical accuracy rests on the unproven assumption that the FRF estimator stays unbiased under the new conditions. A load-bearing error analysis (bias and variance expressions) or Monte-Carlo validation under controlled irregularity is required before the optimality result can be accepted.

minor comments (1)

[Abstract] Abstract: subject-verb agreement error ('The proposed sampling scheme improve' should read 'improves').

Simulated Author's Rebuttal

2 responses · 0 unresolved

Thank you for the opportunity to respond to the referee's comments. We believe the work provides a valuable extension to frequency-domain identification methods, and we address the concerns regarding the estimator properties below.

read point-by-point responses

Referee: [Abstract] Abstract: the nonparametric FRF estimator is asserted to remain unbiased and leakage-free for nonperiodic inputs and nonuniform sampling, yet no derivation, bias expression, or leakage bound is supplied. Standard leakage cancellation relies on periodicity and uniform grids; without an explicit proof that residual bias terms remain independent of sampling irregularity, the subsequent D-optimization of sampling points cannot be guaranteed to improve statistical accuracy.

Authors: The referee correctly identifies that the current manuscript does not include an explicit derivation of the bias and leakage properties for the nonparametric FRF estimator under nonperiodic multisine excitation and nonuniform sampling. We agree that a rigorous proof is necessary to support the claims. In the revised manuscript, we will add a dedicated section providing the bias expression, variance analysis, and a bound on the leakage error that demonstrates independence from sampling irregularity for the given excitation class. This will also justify the D-optimization step. revision: yes
Referee: [Abstract] Abstract: the central claim that the proposed sampling scheme improves statistical accuracy rests on the unproven assumption that the FRF estimator stays unbiased under the new conditions. A load-bearing error analysis (bias and variance expressions) or Monte-Carlo validation under controlled irregularity is required before the optimality result can be accepted.

Authors: We concur that the improvement in statistical accuracy via D-optimal sampling presupposes the unbiasedness of the estimator. To address this, we will incorporate a comprehensive error analysis in the revision, including closed-form bias and variance expressions derived from the nonuniform sampling model. Additionally, we will include Monte-Carlo simulation results under varying degrees of sampling irregularity to empirically validate the unbiasedness and variance reduction. This will strengthen the foundation for the optimality claims. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation remains independent of its fitted outputs

full rationale

The paper introduces a nonparametric FRF estimator for nonperiodic multisine inputs under nonuniform sampling and then optimizes sampling locations via the standard D-optimality criterion from experimental design. No equation or claim reduces the estimator, the leakage bound, or the optimality objective to a quantity that is defined by the same data or by a self-citation chain whose only support is the present work. The central construction therefore does not collapse into a tautology or a fitted-input-called-prediction; the derivation chain is self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based solely on abstract; no explicit free parameters, axioms, or invented entities are stated.

pith-pipeline@v0.9.0 · 5380 in / 876 out tokens · 19806 ms · 2026-05-14T18:57:28.615589+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear
We develop a nonparametric frequency response function estimator suited to such conditions and design irregular sampling schemes that enhance the informativeness of measurements and reduce spectral leakage.
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean alpha_pin_under_high_calibration unclear
Maximizing the determinant of the FIM E[Φ^H_N Φ_N] = A^H R_0 A

Reference graph

Works this paper leans on

24 extracted references · 1 canonical work pages

[1]

and Bernhardsson, B.M

str\"om, K.J. and Bernhardsson, B.M. (2002). Comparison of Riemann and Lebesgue sampling for first order stochastic systems. In Proceedings of the 41st IEEE Conference on Decision and Control, 2002, volume 2, 2011--2016. IEEE

2002
[2]

Bombois, X., Gevers, M., Hildebrand, R., and Solari, G. (2011). Optimal experiment design for open and closed-loop system identification. Communications in Information and Systems, 11(3), 197--224

2011
[3]

and Vandenberghe, L

Boyd, S. and Vandenberghe, L. (2004). Convex Optimization. Cambridge University Press

2004
[4]

Garnier, H., Wang, L., and Young, P.C. (2008). Direct identification of continuous-time models from sampled data: Issues, basic solutions and relevance. In Identification of continuous-time models from sampled data, 1--29. Springer

2008
[5]

(2025 a )

Gonz \'a lez, R.A., Classens, K., Rojas, C.R., Oomen, T., and Hjalmarsson, H. (2025 a ). Finite sample MIMO system identification with multisine excitation: Nonparametric, direct, and two-step parametric estimators. arXiv preprint arXiv:2510.26929

work page arXiv 2025
[6]

(2025 b )

Gonz \'a lez, R.A., Classens, K., Rojas, C.R., Welsh, J.S., and Oomen, T. (2025 b ). Identification of additive continuous-time systems in open and closed loop. Automatica, 173, Art. 112013

2025
[7]

Gonz \'a lez, R.A., Rojas, C.R., Pan, S., and Welsh, J.S. (2021). Consistent identification of continuous-time systems under multisine input signal excitation. Automatica, 133, Art. 109859

2021
[8]

Gonz \'a lez, R.A., van Haren, M., Oomen, T., and Rojas, C.R. (2024). Sampling in parametric and nonparametric system identification: Aliasing , input conditions, and consistency. IEEE Control Systems Letters, 8, 2415--2420

2024
[9]

and Boucher, M.J

Grauer, J.A. and Boucher, M.J. (2020). Aircraft system identification from multisine inputs and frequency responses. Journal of Guidance, Control, and Dynamics, 43(12), 2391--2398

2020
[10]

Hjalmarsson, H. (2005). From experiment design to closed-loop control. Automatica, 41(3), 393--438

2005
[11]

and Johnson, C.R

Horn, R.A. and Johnson, C.R. (2012). Matrix Analysis. Cambridge University Press

2012
[12]

Krantz, S.G. (2001). Function theory of several complex variables, volume 340. American Mathematical Soc

2001
[13]

Ljung, L. (1999). System identification: Theory for the user. PTR Prentice Hall, Upper Saddle River, NJ

1999
[14]

Lundeng rd, K. (2017). Generalized Vandermonde matrices and determinants in electromagnetic compatibility. Ph.D. thesis, M \"a lardalen University

2017
[15]

Miskowicz, M. (2007). Asymptotic effectiveness of the event-based sampling according to the integral criterion. Sensors, 7(1), 16--37

2007
[16]

Morelli, E.A. (2003). Multiple input design for real-time parameter estimation in the frequency domain. IFAC Proceedings Volumes, 36(16), 639--644

2003
[17]

Mu, B., Guo, J., Wang, L.Y., Yin, G., Xu, L., and Zheng, W.X. (2015). Identification of linear continuous-time systems under irregular and random output sampling. Automatica, 60, 100--114

2015
[18]

and Schoukens, J

Pintelon, R. and Schoukens, J. (2012). System identification: A frequency domain approach. John Wiley & Sons

2012
[19]

Pronzato, L. (2008). Optimal experimental design and some related control problems. Automatica, 44(2), 303--325

2008
[20]

Pukelsheim, F. (2006). Optimal design of experiments. SIAM

2006
[21]

and Unbehauen, H

Rao, G.P. and Unbehauen, H. (2006). Identification of continuous-time systems. IEE Proceedings-Control Theory and Applications, 153(2), 185--220

2006
[22]

Schoukens, J., Pintelon, R., and Guillaume, P. (1994). On the advantages of periodic excitation in system identification. IFAC Proceedings Volumes, 27(8), 1115--1120

1994
[23]

Schoukens, J., Rolain, Y., and Pintelon, R. (2006). Analysis of windowing/leakage effects in frequency response function measurements. Automatica, 42(1), 27--38

2006
[24]

Tropp, J.A. (2012). User-friendly tail bounds for sums of random matrices. Foundations of Computational Mathematics, 12(4), 389--434

2012